Mathematically impaired

Most people can’t do simple math. If they could they would have immediately seen something wrong with these results:

A new federal survey about sex and drug use in the United States reveals that an average American man has sex with seven women during his lifetime, compared with four male sexual partners for the average woman.

The latest survey, which claims to have used the high-tech methods to solicit candid answers on sexual activity and illegal drug use, found that approximately 29 percent of men reported having 15 or more female sexual partners in a lifetime, while just over 9 percent of women reported having sex with 15 or more men.

Every time a man has sex with a new partner that woman has sex with a new partner. The writer of this story should have clarified they are not using the usual definition of “average” (the “mean”). If they were then the average for both men and women must be the same. This article clarifies they are using a measure less frequently used by (excuse the pun) lay people, called the “median”.

I don’t have the time to go looking for it in my sex archives but this anomaly in survey results has been known for a long time and it was about five or six years ago they figured it out what was going on.

It turns out prostitutes are under represented in nearly all surveys. Most surveys were done with phone calls during the evening hours. The evening is during the working hours of the “ladies of the night” and hence they are under represented. A simple example will demonstrate why the numbers above, interpreted as a mean, must be bogus and the prostitute answer explains how it could happen.

Suppose there are 100 men and 100 women in a given closed population. Each of the men pair up with one women. But one woman, wanting a little something extra, has sex with not only her partner but the other 99 men as well. The true mean number of partners for the men is (99×2 + 1×1)/100 => 1.99. The true mean number of partners for the women is (99×1 + 1×100)/100 => 1.99. Yet if you did a sample of 20 men and women with a bias against surveying women who were likely to have large numbers of partners then you would probably end up with numbers of 2.0 and 1.0. In any example of heterosexual sex pairings you can come up with in this population the mean number of different partners for men must equal the mean number of partners for the women.